[dpdk-dev] [PATCH] test: fix missing NULL pointer checks

Neil Horman nhorman at tuxdriver.com
Thu Dec 18 22:12:29 CET 2014


On Thu, Dec 18, 2014 at 09:41:47AM +0000, Daniel Mrzyglod wrote:
> In test_sched, we are missing NULL pointer checks after calls to create the 
> mempool and to allocate an mbuf. Add in these checks using VERIFY macros.
> 
> Signed-off-by: Daniel Mrzyglod <danielx.t.mrzyglod at intel.com>
> ---
>  app/test/test_sched.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/app/test/test_sched.c b/app/test/test_sched.c
> index c957d80..9b6e037 100644
> --- a/app/test/test_sched.c
> +++ b/app/test/test_sched.c
> @@ -166,6 +166,7 @@ test_sched(void)
>  	int err;
>  
>  	mp = create_mempool();
> +	VERIFY(mp != NULL,"Error create mempool\n");
>  
>  	port_param.socket = 0;
>  	port_param.rate = (uint64_t) 10000 * 1000 * 1000 / 8;
> @@ -184,6 +185,7 @@ test_sched(void)
>  
>  	for (i = 0; i < 10; i++) {
>  		in_mbufs[i] = rte_pktmbuf_alloc(mp);
> +		VERIFY(in_mbufs[i] != NULL, "Bad packet allocation");
>  		prepare_pkt(in_mbufs[i]);
>  	}
>  
> -- 
> 2.1.0
> 
> 

Looking at the VERIFY macro, its defined as:
#define VERIFY(exp,fmt,args...)                                         \
                if (!(exp)) {                                               \
                        printf(fmt, ##args);
                        return -1;				\
                }

Thats really bad practice, as it embodies a return into the VERIFY macro,
creating hidden function exit points that the programmer can't clean up within.
Every use of the VERIFY macro in test_sched causes the program to return without
freeing any of the memory allocated in the function (not that the function is
any good at cleaning up after itself anyway), but I would recommend that you
modify the macro as such:

#define VERIFY(exp, fmt, args...) \
if (!(exp)) { \
	printf(fmt, ##args); \
	0;\
} else \
	1;\

}


That way you can use the macro like this:

if (VERIFY(in_mbufs[i] != NULL, "Bad packet allocation") {
	//Insert cleanup code here
}

Neil



More information about the dev mailing list